Master's Thesis Award 2025
NAIL Master's Thesis Award 2025
At first glance, topics ranging from greenhouses, prostate cancer, road objects, images and aviation, are seemingly unrelated. However, they have two things in common: they are all master’ thesis topics within AI applications and methods, and they were all finalists for the NAIL Master’s Thesis Award 2025.
The jury and evaluation process
On the 4th of December 2025 we hosted our annual award, celebrating the talented students at NTNU writing Master’ theses within artificial intelligence. This year, in addition to our three usual awards, we teamed up with the Gemini centre MIRA (Medical Imaging Research and AI) for the new MIRA Prize, honouring the best master’s thesis within medical imaging.
A total of 19 theses were nominated, and the jury consisted of selected representatives from NTNU, our partners, and MIRA:
• Chair of Committee: Massimiliano Ruocco, Department of Computer Science
• Giampiero Salvi, Department of Electronic Systems
• Deepti Mishra, Department of Computer Science, NTNU Gjøvik
• Idelfonso Noguiera, Department of Chemical Engineering
• Kristin Jacobsen, partner representative from Arkivverket
• Xiaomeng Su, Department of Computer Science
• Frank Lindseth, representative from MIRA
The jury was tasked with selecting the top three theses within the two categories "AI Applications" and "Theory & Methods", including one winner of the overall best master thesis as well as the MIRA Prize. The jury ended up splitting one of the prizes between two equal nominees, witnessing a high level among the nominated students.
Winners of Master’s Thesis Award 2025
This year the Overall Best Master’s Thesis award went to Oscar Ravik for his thesis: “Integrating Large Language Models with Digital Twins for Autonomous Control”. The jury argued that the work is technically solid, well validated, and demonstrates clear potential for developing AI-driven control systems. Further the awards went to:
Theory/method development: Oskar Jørgensen – “Improving Generative Image Models Using Hybrid Flow Matching and Autoregressive Models”.
Application:
- Olav Finne Præsteng Larsen – “Transfer Learning for Aircraft Landing Trajectory Generation Using Diffusion and Flow Matching Models”
- Sindre Langås Øyen – “Towards Precise Road Object Localization on Smartphones via GNSS-RTK and Visual Sensor Fusion”.
MIRA Prize: Arild Strømsvåg for his thesis “Assessing Self-Supervised Pre-Training with U-Mamba for Prostate Cancer Detection in MRI”.
Congratulations to all the finalists!
A workshop on pitching valuable master’s projects
We continuously work to strengthen academic collaboration with public sector and industry. What better way to inspire our partners than by inviting them to our Master’s Thesis Award, celebrating outstanding master’s theses within AI? Combined with the award ceremony for Best Master's Thesis Award, we hosted a workshop to support our partners on pitching thesis questions to increase likelihood of supervisor endorsement and students selecting their topics. Successful collaboration on master thesis topics brings value to academia, public sector and industry. As part of the workshop, we invited our finalists, supervisors, and partners to participate in panel discussions, offering different perspectives on a master's thesis project and possible collaborations.
Some key takeaways from the panels are:
- For students it is important to have a genuine interest in the project and having an enthusiastic supervisor is contagious.
- For supervisors it is crucial that partners dedicate time and people to support the students, from providing structured and available data to sharing domain expertise.
- Some final advice from our partner SpareBank 1 SMN: Taking the time and laying the groundwork early on is important if you want to succeed in creating effective collaborations in the long run.
The whole event was facilitated and lead seamlessly by PhD Candidate Gro Oleivsgard. We wish to thank her, Massimiliano and the rest of the jury for their efforts. Further we wish to thank all the partners, students and supervisors who participated and for the engagement. We are also happy to announce that the MIRA award will be continued next year together with our three annual awards.
The Mira Prize was sponsored by MIRA, and the remaining awards were sponsored by our partner SpareBank 1 SMN trough the AURA project and was handed out by Executive Director at SpareBank 1 SMN, Astrid Undheim.
2025 Complete list of nominees
- Simen Klemp Wergeland, Magnus Rosvold Farstad - NephroEx: An Explainable AI-Driven Tool for Diuretic Renography Interpretation in Urinary Tract Obstruction
- Espen Bjørge Urheim - Solving Former with Machine Learning Techniques
- Oskar Jørgensen - Improving Generative Image Models Using Hybrid Flow Matching and Autoregressive Models
- Synne Frafjord Moe - Abstractive Summarization for Norwegian Discharge Summaries Using Transformers
- Sondre Sørbye, Jonah Niklas Bjørgsvik Wiecek - BayesGPT: Uncertainy Aware Large Language Models
- Selma Gudmundsen, Thea Sofie Salvesen - Local Explainability to Combat Customer Churn in Sparebank 1 SMN
- Hanna Vodopic - Rethinking Knowledge Management in an AI-Driven World
- A case study on TINE: how "Kupilot" reshapes knowledge management practices
- Oscar Ravik - Integrating Large Language Models with Digital Twins for Autonomous Control
- Mathias Aas and Simen Seeberg-Rommetveit - Domain Adaptation for Norwegian: Integrating Text to Speech, Automatic Speech Recognition, and Large Language Models
- Berk Hadzhamolla - Forecasting Power Transformer Thermal Behaviour Using Neural Ordinary Differential Equations
- Sindre Langås Øyen - Towards Precise Road Object Localization on Smartphones via GNSS-RTK and Visual Sensor Fusion
- Jesper Elverum and Erik Salvesen - Linear-Time Sequence Modeling for Visual Data: A Theoretical and Practical Look at Mamba’s Efficacy on Diverse Medical Imaging Modalities
- Stina Steinsvåg Bogsti - Deep Learning for Automatic Detection of Parkinson's Disease Using Speech Spectral Features and Articulatory Parameters
- Alexander Sandberg - Deploying Trustworthy Deep Reinforcement Learning for Dynamic Positioning: A Live Explainable AI Approach on Real Maritime Cyber- Physical Systems
- Anna Bich-Huyen Doan and Karen Seim Midtlien - Who am I in the Human-AI Symbiosis? How IT students perceive and negotiate their identity as future professionals in the context of GenAI
- Kristoffer N. B. Grude, Sivert A. Eggen, Tord J. Espe - Probabilistic AI for Improved Uncertainty Estimation in Financial Time Series Forecasting: Full Conditional Non-Parametric Distributional Forecasting and Tail Risk Estimation for Stock Returns Using LSTM- and Transformer-based Mixture Density Networks
- Arild Strømsvåg – Assesing Self-Supervised Pre-Training with U-Mamba for Prostate Cancer Detection in MRI
- Andreas Bentzen Winje – Temporal Fusion in Imitation-Based End-to-End Autonomous Driving
- Olav Finne Præsteng Larsen – Transfer Learning for Aircraft Landing Trajectory Generation Using Diffusion and Flow Matching Models